[ome-users] user feedback feedback

[Jason Swedlow]

Alastair-

Thanks again for your email.  My apologies for the delay in  
responding-- have had my head down in a grant and paper.

Harry and Andrea have already responded to most of the details.  i'll  
try to fill in some of the general vision.


On 18 May 2005, at 13:16, Alastair Kerr wrote:

> We have been looking at the web and schoola interfaces for OME  
> 2.4.0 and
> have the following questions. We fully understand that this is very  
> much
> work in progress and we hope that the points raised are of some use to
> the developers.

All feedback is welcome!

>
>
> In General:
>
>
> The concept of the data hierarchy “Projects / datasets /
> images/features” is not entirely compatible with our current structure
> of archiving which tends to be “Date (year)/ Projects/ experiments /
> data sets (including independent replicated data sets by date)/  
> images /
> processed images / attributes (features)”. What is the best way to
> relate the imposed hierarchy to real data acquisition – can a custom
> hierarchy be defined? (this may be somewhat related to the Classifier
> functionality but we are not clear how exactly to implement it?)
>

We have imposed the P/D/I hierarchy within OME.  But with the release  
of 2.4.0, we have supported more flexible hierarchies, totally  
defined by the user.  These are still limited to a Category Group  
(e.g., cell cycle position), and Categories within that (e.g.,  
prophase, metaphase, etc.).  There are some serious limitations to  
these:

1) The number of levels in a user-defined hierarchy is fixed: there  
is a CategoryGroup, and a Category.  That's it.
2) An Image can belong to one and only one Category within a given  
CategoryGroup, but it can belong to multiple CategoryGroups.

These are now supported in the server and UI software, and we want to  
determine whether our current implementation is useful and/or whether  
they present a severe hindrance to users.  Sounds like you want more  
flexibility


>
> Is OME able to import an existing directory structure (e.g. from an
> archived DVD) and use this to automatically define projects /  
> datasets,
> perhaps through an interactive dialogue? Moreover can you allow
> importing of images from nested directories by selecting only the top
> level directory during image import (we tried it but failed to get
> correct import).

Currently, the OME web UI supports importing a single directory, but  
not a whole directory structure.  The functionality here needs to be  
updated (its actually pretty old).  Can you please note this on  
http://bugs.openmicroscopy.org.uk/  We use this facility for bugs,  
but also for feature requests.  Using this method makkes all requests  
public, and also keeps you notified of discussion and progress.

>
>
> Related to the above – we remember that utilising an existing  
> directory
> structure was something that the “Catalyser” software could do quite
> well. Is the Catalyser software still under consideration as a  
> front end
> for OME in this regard?
>

You will have to talk to the Axiope people on this-- we have  
discussed this with them, but its really up to them and their  
priorities.  We would love it.

>
> Currently it is unclear that while the Web interface is referred to as
> “light” it seems to have more functionality than the Java interface?
> Also we are not clear on the logic of the division of functionality
> between the Java and web client – ultimately will all tasks (such as
> image import, chain construction and execution of chains) be available
> from the Schoola interface?
>

Yes, both clients will become fully developed.  As you noted, they  
are both works in progress.  Ours goals for Shoola was initially to  
provide reasonable functionality for a 5D viewer and to explore what  
the possibilities for client/server communication while analysing an  
Image.  The functionality in Shoola and the Web UI are certainly  
different-- this is deliberate, as we are trying too leverage the  
advantages of both environments.  we do have users who just want a  
web browser-based system, others who want a much richer client.  For  
example, you can't yet import data within Shoola (due to some  
problems we have in the server system, which we are fixing).  The  
Viewer on Shoola is more sophisticated, and you can directly measure  
volumes, intensities etc on Shoola.  On the Web UI, there is much  
better support for running large-scale analyses, and viewing the  
results. See http://www.openmicroscopy.org/getting-started/ for some  
screenshots.

In general however, we'd love to hear what you want from clients, and  
what they should look like.,  Some of our users don't give two hots  
for quantitative analysis, and just want a data manager.  Others are  
much more concerned with linking results from quantitative analyses  
to images.


>
> Will any “wizards” be available as a pull down on the main menu bar  
> when
> you first open the program to step new users through common tasks?  
> E.g.
> defining of a project and data set followed by import of data;
> extracting / modifying and adding image metadata; locating and  
> viewing a
> data set. This is perhaps the largest barrier to new users starting
> using OME as the options and way to proceed are not always intuitive.

We severely lack all the nice things that users have come too expect  
from fully-developed software-- wizards, menus, tool-tips, help,  
etc.  Is just a matter of resources-- specifically pairs of hands.   
We are trying to fix some of this summer, and will continue to work  
on this.  Any and all help is much appreciated!!!!!

>
>
> Is there a list of data formats that can be read by OME as well as the
> list of programs that can import OME XML format? At the moment we can
> only ascertain whether a file format is compatible by reading the  
> import
> process notes. (For example are we correct in assuming that .avi moves
> and Perkin Elmer Ultraview format are not supported).
>

The current list is at http://www.openmicroscopy.org/getting-started/ 
import.html  (Yes, we need to link out to the web page).  In general,  
we are not yet supporting compressed formats like JPEG, MPEG and AVI,  
because of the data-loss problems there formats cause.  Tell us what  
you think of this policy.


>
> ---------------------------------------------------------------------- 
> -------------------------------------
>
>
>
> Java Client [using stable relase 2.4.0]:
>
>
> Currently the “help” offers no content. It might be useful to replace
> this with a hyperlink to the online “Getting started tutorial”  
> resource,
> that we find quite helpful, as a stop-gap to any planned expansion of
> this functionality.

Good point.

>
>
> Particularly in the case of the Java client, both the functionality
> available and how to progress through operations is not intuitive.  
> This
> is complicated by the distribution of functionality between the  
> Java and
> Web interfaces.

Exactly-- they are complementary, and only barely over-lapping.   
Basically, with the limited resources we have, we are trying to  
explore as much as possible.  But the expected workslows in both UIs  
are not intuitive too many users, as we don't have alot of HCI  
experience on the project.  However, our main priority right now is  
to fix some of the problems with the system our clients and users  
have uncovered.  That will work will go on through this summer.  But  
we will work on these suggestions later this year.  In addition, we  
have submitted grant applications to increase our staff and allow us  
to dedicate more time and effort to usability and user support issues.

A general point: at any time, we have a Roadmap that runs about six  
months into the future, with all of developer resources dedicated.   
User/developer requests can be added to t his, and of course we  
devote resources to dealing with bugs, etc.  But anyone out there who  
wants to help otu is more than welcome.  This is a huge task-- we are  
building tools that have never been built before.  Real help--  
skilled developers dedicated to working with the team-- are always  
welcome!!!


>
> Would it be possible to provide a working chains demo? I.e.  
> downloadable
> datasets coupled to a library of basic chains with which they will  
> work,
> to get people started? The current chain examples are not entirely  
> clear
> to us and possibly won’t even work with the data sets we are using.
>

I think Harry responded to this-- good idea

> There is clearly a great deal of functionality, however, much seems to
> be “hidden” under “cryptic” icons or names, or requires the user to  
> make
> selection by direct or right mouse (ctrl+Mouse) in places other  
> than the
> intuitive locations and also having to confirm a requested action in
> some cases by clicking another “button”. In this last case perhaps a
> popup dialogue box asking to “continue or add/do another” might lead
> users on better. By random clicking and reference to the helpful  
> online
> “Getting started tutorial” we could muddle through eventually, but his
> does not seem like it would be appealing to primary Users (Post  
> Docs and
> PhD’s.


Again, all great.

>
> It would be good to have more (all relevant?) functions duplicated in
> categorised pull down menus on the relevant windows – if I can’t  
> find a
> function easily I find it useful to just browse the contents of pull
> down menus to give me a hint what is available rather than browsing  
> the
> individual icon buttons.
>
> Separate “editing”, “viewing” and “processing” workflows are  
> defined but
> I still find I keep accidentally skipping between them. Could they be
> even better defined and separated out? This might be sorted out by
> Wizards to get you used to the separate paths or by grouping the  
> related
> functionalities (editing/ viewing/ processing) under appropriately
> labelled pull down menus.
>
> Could the viewing of multi wavelength data be available in the image
> view directly as well as under rendering?


All good ideas.

>
> Chains and chain builder do not fully convey to me what is  
> intended. The
> way of putting chains together is fairly intuitive however the  
> amount of
> info that pops up immediately is overwhelming to a novice user. The
> navigator function (overview, a good idea) seems to be only
> intermittently operational in our hands which causes confusion. It  
> looks
> like you should be able to draw a rectangle (red area) round the  
> window
> area you want to see from the overview but this does not always  
> work. It
> might be better if it could be expanded to a size where the text is
> readable – or the full text replaced by a “title” for that section.
> “Mouse-over” to highlight areas and the way mouse clicking in  
> different
> locations to expand and more contents within a window appears clumsy –
> could simply clicking the section you want and having a fit to a  
> window
> which you could drag to size/shape not work better?
>
> The ROI tool does not seem to allow local manipulation / analysis  
> on the
> part of the image defined by the ROI – would be good to be able to
> resave the subregion or limit processing to that region.
> Tried to define
> a subregion of an image stack to do subsequent processing to reduce  
> the
> processing time. Also tried to view local region to play as a movie  
> – to
> speed up the player (in our hands the possible play speed appeared  
> to be
> severely limited by the size of image, certainly well below the  
> selected
> play speed).

Better support for sub-region definition and viewing will come out of  
work we are doing now to fix up the OME Image Server-- stay tuned!!!!

Yes, very much a work in progress.  Let us know if it is worth  
getting together.  We are considering a meeting either here in the  
UK, or possibly in the US, tied to one of the scientific meeting  
(ASCB???).  If we did the latter, would people come??

Again, many thanks!!!!

Cheers,

Jason


**************************
MSI/WTB Complex
The University of Dundee
Dow Street
Dundee  DD1 5EH
United Kingdom

phone (01382) 345819
Intl phone:  44 1382 345819
FAX   (01382) 348072
email: jason at lifesci.dundee.ac.uk

Lab Page: http://www.dundee.ac.uk/lifesciences/swedlow/
Open Microscopy Environment: http://openmicroscopy.org
**************************